How Easy is it to Learn a Controlled Natural Language for Building a Knowledge Base?

TL;DR

This study investigates how both novices and experts learn and use a controlled natural language, OWL Simplified English, for building knowledge bases, highlighting the skills needed and challenges faced during the process.

Contribution

It provides an exploratory analysis of user behaviors and difficulties in learning a CNL for knowledge engineering without guidance, using eye-tracking and screen recordings.

Findings

Identified key skills required for learning CNL

Observed common difficulties faced by users

Provided insights into user behavior during knowledge base construction

Abstract

Recent developments in controlled natural language editors for knowledge engineering (KE) have given rise to expectations that they will make KE tasks more accessible and perhaps even enable non-engineers to build knowledge bases. This exploratory research focussed on novices and experts in knowledge engineering during their attempts to learn a controlled natural language (CNL) known as OWL Simplified English and use it to build a small knowledge base. Participants' behaviours during the task were observed through eye-tracking and screen recordings. This was an attempt at a more ambitious user study than in previous research because we used a naturally occurring text as the source of domain knowledge, and left them without guidance on which information to select, or how to encode it. We have identified a number of skills (competencies) required for this difficult task and key problems that authors face.